Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ixen20

Download by: [Australian Catholic University] Date: 27 July 2017, At: 01:45

Xenobiotica

the fate of foreign compounds in biological systems

ISSN: 0049-8254 (Print) 1366-5928 (Online) Journal homepage: http://www.tandfonline.com/loi/ixen20

The metabolism and drug-drug interaction

potential of the selective prostacyclin receptor

agonist selexipag

Carmela Gnerre, Jérôme Segrestaa, Swen Seeland, Päivi Äänismaa, Thomas

Pfeifer, Stephane Delahaye, Ruben de Kanter, Tomohiko Ichikawa, Tetsuhiro

Yamada & Alexander Treiber

To cite this article: Carmela Gnerre, Jérôme Segrestaa, Swen Seeland, Päivi Äänismaa, Thomas

Pfeifer, Stephane Delahaye, Ruben de Kanter, Tomohiko Ichikawa, Tetsuhiro Yamada & Alexander

Treiber (2017): The metabolism and drug-drug interaction potential of the selective prostacyclin

receptor agonist selexipag, Xenobiotica, DOI: 10.1080/00498254.2017.1357088

To link to this article: http://dx.doi.org/10.1080/00498254.2017.1357088

Accepted author version posted online: 24

Jul 2017.

Submit your article to this journal

Article views: 4

View related articles

View Crossmark data

The metabolism and drug-drug interaction potential of the selective

prostacyclin receptor agonist selexipag

Carmela Gnerre

, Jérôme Segrestaa

, Swen Seeland

, Päivi Äänismaa

, Thomas Pfeifer

Stephane Delahaye

, Ruben de Kanter

, Tomohiko Ichikawa

, Tetsuhiro Yamada

and

Alexander Treiber

Preclinical Pharmacokinetics and Metabolism, Actelion Pharmaceuticals Ltd, Allschwil, and

Pharmacokinetics and Safety Assessment Dept., Nippon Shinyaku Co., Ltd, Kyoto, Japan

Current address: Märtktweg 30, D-79576 Weil am Rhein

Corresponding Author: Carmela Gnerre, Idorsia Pharmaceuticals Ltd, Hegenheimermattweg 91, CH-4123

Allschwil, Switzerland. Email: carmela.gnerre@idorsia.com

Acknowledgments

The authors would like to thank Noura Akel, Julia Friedrich, Joëlle Pichot, Ervan Hauy and

Ali Selimi for their dedication and experimental contributions and Hashem Salloukh and

Alexey Veligodskiy for their assistance in the preparation of this manuscript.

Declaration of interest

This work was funded by Actelion Pharmaceuticals Ltd, Allschwil, Switzerland and Nippon

Shinyaku Co, Ltd, Kyoto, Japan. The authors report no declarations of interest.

All experiments described in this report have been conducted in the research facilities of

Actelion Pharmaceuticals Ltd and Nippon Shinyaku. During manuscript review, Actelion

Pharmaceuticals Ltd was acquired by Johnson & Johnson, and its drug discovery and early

development activities subsequently transferred into a newly created company, Idorsia

Pharmaceuticals Ltd. With the exception of Thomas Pfeifer, Tetsuhiro Yamada and

Tomohiko Ichikawa, all authors of this report were employees of Idorsia Pharmaceuticals Ltd

at the time of manuscript publication.

Accepted Manuscript

The metabolism and drug-drug interaction potential of the selective

prostacyclin receptor agonist selexipag

Abstract

- The metabolism of selexipag has been studied in vivo in man and the main excreted

metabolites were identified. Also, metabolites circulating in human plasma have been

structurally identified and quantified.

- The main metabolic pathway of selexipag in man is the formation of the active metabolite

ACT-333679. Other metabolic pathways include oxidation and dealkylation reactions. All

primary metabolites undergo subsequent hydrolysis of the sulphonamide moiety to their

corresponding acids. ACT-333679 undergoes conjugation with glucuronic acid and aromatic

hydroxylation to P10, the main metabolite detected in human faeces.

- The formation of the active metabolite ACT-333679 is catalysed by carboxylesterases,

while the oxidation and dealkylation reactions are metabolized by CYP2C8 and CYP3A4.

CYP2C8 is the only P450 isoform catalyzing the aromatic hydroxylation to P10. CYP2C8

together with CYP3A4 are also involved in the formation of several minor ACT-333679

metabolites. UGT1A3 and UGT2B7 catalyze the glucuronidation of ACT-333679.

- The potential of selexipag to inhibit or induce cytochrome P450 enzymes or drug transport

proteins was studied in vitro. Selexipag is an inhibitor of CYP2C8 and CYP2C9 and induces

CYP3A4 and CYP2C9 in vitro. Also, selexipag inhibits the transporters OATP1B1,

OATP1B3, OAT1, OAT3 and BCRP. However, due to its low dose and relatively low

unbound exposure, selexipag has a low potential for causing drug-drug interactions.

Accepted Manuscript

Word Count:

Abstract (213), Introduction (506), Materials and Methods (5238), Results (1761) Discussion

(1057), References (722), Table legends (291) Total (9788)

Key words: Prostacyclin receptor agonist, selexipag, metabolism in man, drug

interaction, cytochrome P450, drug transporter

Accepted Manuscript

Introduction

Pulmonary arterial hypertension (PAH) is a life-threatening disease characterized by

endothelial dysfunction, impaired signalling and vascular remodelling, leading to right

ventricular dysfunction and eventually failure if left untreated (Galiè et al., 2015; Simoneau

et al., 2013). In PAH, there is increased production of vasoconstrictor mediators such as

endothelin, and reduced production of the vasodilator mediators prostacyclin and nitric oxide

(Clapp and Gurung, 2015; Tuder et al., 1999). Approved PAH-specific pulmonary vasoactive

therapies target three major pathophysiological pathways, i.e. the nitric oxide, endothelin-1

and prostacyclin pathways. They comprise a soluble guanylate cyclase stimulator,

phosphodiesterase-5 inhibitors, endothelin receptor antagonists and prostanoids analogues

(Humbert and Ghofrani, 2016; Gailè et al., 2013; Taichman et al., 2014). Selexipag is an

orally active, non-prostanoid, selective prostacyclin receptor agonist, which is rapidly

hydrolysed to its active metabolite ACT-333679. Selexipag (Uptravi

) was approved recently

in several countries/regions including the USA and Europe based on results of the large,

randomized, placebo-controlled, event-driven clinical study GRIPHON. In that study,

selexipag when given to PAH patients significantly reduced the risk of the primary composite

endpoint of all-cause death or a complication related to PAH versus placebo (Sardana et al.,

2016; Duggan et al., 2017).

The approved treatment of selexipag is based on an individual uptitration scheme. The

starting dose is 200 μg twice daily, with increases of 200 μg twice daily every week up to the

highest tolerated dose of 1600 μg twice daily (Hardin and Chin, 2016). Selexipag is rapidly

absorbed, with maximum plasma concentration achieved within 2 hours post-dose. Plasma

exposure to ACT-333679 is approximately 4 times higher than that to selexipag (Bruderer et

al., 2014). The oral bioavailability of selexipag is 49% (Kaufmann et al., 2017). It has been

demonstrated that ACT-333679 is a 37-fold more potent, selective agonist at human IP

Accepted Manuscript

receptors than selexipag (Gatfield et al., 2017). This potency difference, together with the

higher unbound abundance of ACT-333679 in plasma as compared to selexipag, makes it the

most relevant pharmacological entity.

Selexipag and ACT-333679 are lipophilic molecules with logD values of 2.1 and 2.2,

respectively. Both compounds are negatively charged at physiological pH and highly bound

to plasma proteins.

Here, we describe the metabolism and the drug-drug interaction potential of selexipag in

man. Radiolabelled selexipag was given orally to healthy volunteers, and plasma and excreta

collected for up to 7 days. The chemical structures of the most prominent faecal metabolites

were determined but metabolite structures could not be identified in plasma and urine due to

the low absolute radioactivity concentrations. To further understand selexipag's metabolic

pathways, plasma samples from a clinical phase I study (Bruderer et al., 2014) were used to

determine the chemical nature and quantity of circulating metabolites. Also, metabolism of

radiolabelled selexipag and ACT-333679 was studied in vitro with human plasma and liver

preparations. Metabolic patterns were recorded using high performance liquid

chromatography (HPLC) coupled to radiodetection and metabolites structures were identified

using mass spectrometry (MS) technology and authentic references. The potential of both

compounds to inhibit or induce cytochrome P450 enzymes or drug transporters was

investigated in vitro.

Accepted Manuscript

Material and Methods

Chemicals and Reagents

[

C]-Selexipag (ACT-293987) and its metabolite [

C]-ACT-333679 (Figure 1) were

manufactured at Nippon Shinyaku (Kyoto, Japan). The specific activities of [

C]-selexipag

and [

C]-ACT-333679 were 56.2 and 56.9 mCi/mmol, respectively. The radiochemical

purity of both compounds was 98% or above. Non-labelled selexipag (lot 32 or 25) and ACT-

333679 (lot 10) were manufactured at Nippon Shinyaku with a purity of 99.7 % or higher.

The deuterated standards ACT-293987B (batch 2122-DB/70) and ACT-333679B (batch

2122-AA/60) used for bioanalytical purposes were synthesized at Actelion Pharmaceuticals

Ltd (Allschwil, Switzerland). Metabolites P4, P10, P11, P12, P13, P14, P18, P34 and P35

were synthesized at Nippon Shinyaku. All other chemicals and solvents used throughout this

study were of highest commercially available quality and obtained from Sigma-Aldrich,

Fluka (Buchs, Switzerland), American Radiolabeled Chemicals (St. Louis, USA), NEN

Radiochemicals Perkin Elmer (Boston, USA), Nacalai Tesque (Kyoto, Japan) or Wako

Chemicals (Osaka, Japan). Liquid scintillation cocktails, Emulsifier Scintillator Plus, Hionic-

Fluor Irga Safe Plus and Optiflow Safe 2, were from Perkin Elmer (Zürich, Switzerland or

Waltham, USA) and Berthold Technologies (Regensdorf, Switzerland). Recombinant

cytochrome P450 and UGT enzymes were purchased from Cypex Ltd (Dundee, United

Kingdom) or BD Biosciences (Japan). Pooled human liver microsomes were obtained from

BD Biosciences (Basel, Switzerland or Tokyo, Japan), Xenotech (Kansas City, USA) or

Biopredic International (Rennes, France), and fresh human hepatocytes were obtained from

Cytonet (Hannover, Germany).

Accepted Manuscript

Identification of metabolites in human faeces

Oral dosing with [

C]-selexipag in human and preparation of faecal extracts

A single-centre, open-label, non-randomized study after single dose administration of 400 µg

selexipag and a radioactive dose of 45 µCi was carried out in 6 subjects (Kaufmann et al.,

2012). This study was conducted in compliance with the principles of the Declaration of

Helsinki. The study protocol was approved by the Edinburgh Independent Ethics Committee

for Medical Research. The faeces collected from that study at time intervals of 24 h up to 168

h post-dose were homogenized in an approximately equal weight of water/acetone. Faecal

homogenate samples from each time interval were twice centrifuged at 50 g and metabolites

were extracted after acidification with 1M hydrochloric acid with ethyl acetate. After mixing

for 10 minutes, the samples were further centrifuged at 2160 g for 10 min.

Radioactive metabolite profiling

Radioactivity profiling was performed using an Agilent 1100 HPLC system with an L-

column ODS (5 μm; 150 × 4.6 mm ID; Chemicals Evaluation and Research Institute, Tokyo,

Japan) coupled in-line with a PerkinElmer D-515TR radiodetector. The mobile phase was

acetonitrile/10 mM ammonium acetate, pH 4.95 (10:90, v/v) as phase A and

acetonitrile/10 mM ammonium acetate, pH 4.95 (90:10, v/v) as phase B. The HPLC flow rate

was 1 mL/min and the column temperature was 40°C. The radiodetector utilized a 500-μL

liquid cell with a scintillant (Ultima Flo AP) flow rate of 2 mL/min. The drug and metabolites

were eluted with the following gradient program: starting from 20% B to 70% B at 30 min,

then to 100% B at 30.01 min and maintained until 33.01 min; then back to 20% B at 33.01

min. Radiochromatograms generated from the radiometry were processed using ChemStation

(Agilent) and the Flo-One software package (version 3.6, PerkinElmer).

Accepted Manuscript

Metabolite identification

The structures of main metabolites in human faecal samples were elucidated using

LC/radiodetection and mass spectrometry in comparison with synthetic standards. In

addition, the results of incubation of [

C]ACT-333679 with recombinant CYP2C8 (Nosan

Corporation (Kanagawa, Japan) were used to support the identification and structural

elucidation of those metabolites. The incubation mixture contained 75 μg/mL CYP2C8

recombinant protein, 20 μg/mL [

C]ACT-333679, 2.7 μCi/mL, 1 mM NADPH, 3.3 mM

MgCl

, and 100 mM potassium phosphate buffer (pH 7.4). Non-labelled ACT-333679 was

incubated using the same procedure. Reactions were initiated by the addition of NADPH and

incubated at 37 °C. After 120 min, reactions were terminated with 2.5 mL of methanol. The

mixtures were centrifuged at 2160 g for 10 min at room temperature and the supernatants

were subjected to HPLC analysis. For LC/radiochromatography, an Agilent 1100 HPLC

system coupled in-line with a 515TR radiodetector was used. To isolate a peak with a

retention time identical to that of the main metabolite in faecal samples, the supernatant was

fractionated by chromatography on an L-column ODS with methanol-0.1% formic acid

(65:35, v/v). The isolated peak was identified by comparison of its HPLC retention time and

mass fragmentation pattern with the one of an authentic standard. The identified metabolite

was further incubated with recombinant human CYP2C8 protein as described previously. The

resulting metabolite peak, whose retention time was identical to that of the second most

abundant in faecal samples, was isolated by HPLC and identified by comparison of its HPLC

retention time and mass fragmentation pattern with the one of an authentic standard.

Chromatographic separation of metabolites was achieved using reversed-phase LC (an L-

column ODS, 5 μm, 150 × 4.6 mm ID, for LC/radiochromatography analysis and an L-

column ODS, 3 μm, 100 × 2.1 mm ID, for LC-MS/MS analysis). Phase A was

acetonitrile/10 mM ammonium acetate, pH 4.95 (10:90, v/v), phase B acetonitrile/10 mM

Accepted Manuscript

ammonium acetate, pH 4.95 (90:10, v/v). For LC/radiochromatography analysis, the

metabolites were eluted with the following linear gradient: equilibration at 20% B; 0 to 30

min, 20 to 70% B; 30 to 30.01 min, 70 to 100% B; 30.01 to 33 min, 100% B; and 33 to 33.01

min, 100 to 20% B. The flow rate was 1.0 mL/min in LC/radiochromatography analysis.

Radioactive metabolites were detected by mixing the eluent with Ultima Flo AP liquid

scintillant (3 mL/min) and passing it through a detection cell (500 μL) in the radiodetector.

The acquisition time was 30 min. Radiochromatograms generated from the radiometry were

processed using the FLO-ONE software package (version 3.6).For LC-MS/MS analysis , an

Acquity UPLC system with an L-column ODS (2.1 × 100 mm, 3 μm, Chemicals Evaluation

and Research Institute, Tokyo, Japan) coupled with an Acquity PDA UV detector and a

Waters Quattro Micro API mass spectrometer. The mobile phase was acetonitrile/10 mM

ammonium acetate, pH 4.95 (10:90, v/v) as solvent A and acetonitrile/10 mM ammonium

acetate, pH 4.95 (90:10, v/v) as solvent B. The flow rate was 0.3 mL/min and the column

temperature was 40°C. The drug and metabolites were eluted with the following gradient

program: starting from 20% B to 100% B at 10 min and maintained until 12 min; then back to

20% B at 12.01 min. The optimal tune parameters for the detection of selexipag and its

metabolites were as follows: capillary 3.5 kV; cone 30 V; extractor 2 V; source 120 °C; and

desolvation 400 °C. The scan range was set from 400 to 500. The MS data were processed

with MassLynx software (version 4.1, Waters).

Identification and quantification of metabolites circulating in human plasma

Human plasma samples from the 1800 µg b.i.d. and the 1600 µg b.i.d. dose group of a

clinical phase I multiple ascending dose study (Bruderer et al., 2014) were used to determine

the chemical structure and the amount of circulating metabolites, respectively. Once the

structural elucidation was performed, the metabolite quantification was subsequently done

Accepted Manuscript

with available synthetic references. Plasma, acidified with 1M HCl upon original preparation,

was pooled per time-point until 12 h post-dose and precipitated with acetonitrile containing

the internal standards ACT-293987B and ACT-333679B. Metabolites were investigated

using liquid chromatography combined with high resolution mass spectrometry and

radiodetection (split ratio MS/radiodetector, 1/5). The LC-MS system for metabolite

identification consisted of a Rheos Allegro (Flux Instruments) and a LC-10AD VP or LC-

20ADXR VP (Shimadzu) pump, an HTC PAL (CTC Analytics) autosampler and a Hot Dog

200/300 column oven (Thermo Fisher). Chromatographic separation was achieved using two

Gemini RP 150 x 2 mm columns with 3 µm particle size, in serial connection at 45°C. Eluent

A consisted of water containing 10 mM ammonium formate at pH 4.2, adjusted with formic

acid and eluent B consisted of acetonitrile with 0.1% formic acid. The flow rate was 0.2

mL/min. The following gradient was applied: starting from 20% B to 35% B at 5 min, then to

40% B at 15 min, to 60% B at 40 min, to 75 B at 47 min, to 95% B at 57 min maintained

until 58 min; then back to 20% B at 59 min. The HPLC was coupled with an LTQ Orbitrap

Discovery mass spectrometer equipped with a heated electrospray source H-ESI (Thermo

Fisher, San Jose, CA, USA) and operated under the following conditions: ESI potential 3 kV,

capillary temperature 270°C, vaporizer temperature 200°C, capillary 20.5 V, tube lens 100 V.

Sheath and auxiliary gas (nitrogen for both) were supplied using 60 and 5 arbitrary units

respectively. Full scan high resolution LC-MS chromatograms were recorded in positive ion

mode at a resolving power of 30000. Full scan high resolution LC-MS chromatograms were

generated to identify metabolites using the Metworks 1.2 and Xcalibur 2.0.7 software. The

determined exact mass error of all metabolites was below 3 ppm. Putative metabolites were

further investigated by MS/MS experiments in positive ion mode using collision induced

dissociation in the ion trap (CID) with 40% relative collision energy and high energy

collision dissociation (HCD) in the collision cell at 40 eV normalized collision energy as

Accepted Manuscript

fragmentation tools. On-line H/D exchange experiments were performed to investigate the

number of acidic protons for all metabolites using D

O instead of the water in eluent A. The

quantification of metabolites was carried out using an API5000 (AB Sciex Instruments,

Toronto, Canada) triple-stage quadrupole mass spectrometer equipped with a Turbo ionspray

interface operating in positive ion mode using qualified methods, calibration curves and

quality control samples.

Metabolism of selexipag and ACT-333679 in vitro

Hydrolysis of selexipag to ACT-333679

[

C]-selexipag was incubated at a final concentration of 10 µM in a 100 mM phosphate

buffer (pH=7.4, 37 °C) with pooled human liver microsomes (Xenotech, Kansas City, USA)

at a protein concentration of 1 mg/mL proteins. The reaction solutions without [

C]-

selexipag were pre-incubated at 37ºC for 5 min. The reaction was initiated by addition of

[

C]-selexipag and incubation was continued for 60 min at 37°C. Preliminary experiments

showed that the formation of [

C]-ACT-333679 from [

C]-selexipag was linear up to 60

min and 1 mg/mL of protein. The reactions were terminated by addition of one volume

equivalent of acetonitrile, and the samples were centrifuged at 16060 g for 5 min. An aliquot

of the supernatant was submitted to thin layer chromatography (TLC) analysis. Control

experiments in which acetonitrile was added to the reaction prior to [

C]-selexipag were run

under otherwise identical conditions to verify that ACT-333679 formation was indeed

enzyme-dependent.

In the inhibition experiments, carboxylesterase (CES) inhibitors bis(p-nitrophenyl) phosphate

(BNPP) and phenylmethylsulfonyl fluoride (PMSF), paraoxonase inhibitor

ethylenediaminetetraacetic acid (EDTA), and acetylcholine esterase inhibitor eserine were

added to the reaction mixture prior to initiating the reaction, each at concentrations of 0.01,

Accepted Manuscript

0.1 and 1 mM. [

C]-ACT-333679 produced in the reaction was determined by TLC. Silica

gel 60 F254, 0.25 mm (Merck, Darmstadt, Germany) was used as a TLC plate and a mixture

of dichloromethane, methanol and formic acid (10/2/0.1, v/v/v) as the developing solvent.

After development, the TLC plate was covered with a protective film (Saran Wrap, Asahi

Kasei Home Products, Tokyo, Japan). The TLC plate was in contact with an imaging plate

(TYPE BAS-MS, Fujifilm) and exposed in the lead shield box (BAS-SHB 2040, Fujifilm).

After exposure, image analysis was performed using a bio-imaging analyzer (BAS-2500) to

fractionate [

C]-selexipag, [

C]-ACT-333679 and other parts. The [

C]-ACT-333679

fraction (%)of the total radioactivity was used to calculate the formation rates of ACT-

333679 in the presence and in the absence of inhibitors. Finally, the remaining activity (%) in

the presence of inhibitors was calculated based on those rates.

Selexipag incubations with human liver preparations and plasma

[

C]-selexipag was incubated at 37 °C with pooled human liver microsomes (Biopredic

International, Rennes, France) containing 3 mg/mL proteins, in a 100 mM phosphate buffer

(pH=7.4) at a final concentration of 10 µM. The reaction was initiated by addition of the pre-

warmed NADPH-regenerating system and incubation was continued for 60 min at 37°C. In

the inhibition experiments with selexipag and sulfaphenazole (3 µM, CYP2C9), montelukast

(3 µM, CYP2C8), N-benzylnirvanol (5 µM, CYP2C19), quinidine (1 µM, CYP2D6), and

ketoconazole (1 µM, CYP3A4), an aliquot of concentrated inhibitor stock solution was added

to the reaction mixture prior to the initiation of the reaction. In the inhibition experiments

with the mechanism-based inhibitors azamulin (5 µM, CYP3A4), furafylline (20 µM,

CYP1A2) and ticlopidine (30 µM, CYP2B6), microsomes were first incubated for a period of

10 min at 37°C in the presence of the NADPH-regenerating system. The NADPH-

regenerating system was prepared as a 10-fold concentrated stock solution and kept at 20

Accepted Manuscript

°C. It consisted of 11 mM NADP, 100 mM glucose-6-phosphate and 50 mM magnesium

chloride in 100 mM phosphate buffer (pH 7.4). 20 UI/mL of glucose-6-phosphate

dehydrogenase were added in this mixture immediately before use. [

C]-selexipag was then

added at the end of the pre-incubation period and the reaction continued for another 60 min.

All experiments were performed in parallel in the presence or absence of 200 µM of the

carboxylesterase inhibitor BNPP (n=1). The organic solvent in all incubations was kept

below 1 % (v/v). The reactions were terminated by addition ice-cold acetonitrile, and the

samples were centrifuged at 20800 g for 10 min and 10 °C. An aliquot of the supernatant was

submitted to HPLC analysis. Control incubations in the absence of the NADPH regenerating

system or liver microsomes were performed in parallel under otherwise identical conditions.

Fresh human hepatocytes were provided in ready-plated cultures by Cytonet (Hannover,

Germany) and used immediately after receipt. They were cultured in fresh William's E

medium supplemented with 10 % foetal calf serum (v/v), 0.7 μM insulin, 10000 UI/mL

penicillin and 10 mg/mL streptomycin for 3 h prior to use according to the protocols provided

by the suppliers. At the end of the pre-incubation period, the medium was removed from each

well and replaced by pre-warmed (37 °C) William's E medium supplemented as described

above and additionally fortified with 1.0 μM hydrocortisone, 400 μM glutamine, and [

C]-

selexipag or [

C]-ACT-333679 at final concentrations of 10 μM. Triplicate wells were

sampled after up to 24 h of incubation by addition of methanol containing 0.1 M hydrochloric

acid and the entire well content was transferred into cryo-vials. Samples were stored frozen at

-20°C pending analysis.

Human plasma containing EDTA and prepared at Actelion Pharmaceuticals Ltd (Allschwil,

Switzerland) was fortified with [

C]-selexipag at a final concentration of 10 μM. After an

incubation time of up to 6 h, plasma proteins were precipitated by addition of an ice-cold

Accepted Manuscript

mixture of acetonitrile and methanol. The samples were centrifuged at 20800 g for 10 min at

10 °C, and a 450 μL-aliquot of the supernatant was submitted to HPLC analysis.

Selexipag incubations with recombinant cytochrome P450 enzymes

Incubations of [

C]-selexipag with recombinant human P450 enzymes CYP1A2, 2B6, 2C8,

2C9, 2C19, 2D6 and CYP3A4 were performed at a single concentration of 10 µM. An aliquot

of the [

C]-selexipag stock solution was added to a 100 mM phosphate buffer (pH 7.4, 37

°C) containing the P450 enzyme at a concentration of 100 pmol/mL. A co-incubation of

recombinant CYP1A2 and CYP3A4 was also performed and contained each enzyme at a

concentration of 100 pmol/mL under otherwise same experimental conditions. The reaction

was initiated by adding pre-warmed NADPH-regenerating system and incubation was

continued for 60 min at 37°C in an Eppendorf thermomixer at 650 rpm. The reaction was

terminated by addition of one volume equivalent of ice-cold acetonitrile. The samples were

centrifuged at 20800 g for 10 min and 10°C. A 100 µL aliquot of the supernatant was

submitted to HPLC analysis. Control incubations in the absence of the NADPH-regenerating

system, as well as incubations with control bactosomes without P450 enzymes were

performed in parallel under otherwise identical conditions (n = 1).

Analytical method for selexipag incubations

Metabolic profiles from incubations with human liver fractions and recombinant enzymes

were analysed using HPLC coupled with radioactivity detection. The analytical system

consisted of two Shimadzu HPLC pumps LC-30AD (Shimadzu, Reinach, Switzerland)

equipped with a Shimadzu membrane degasser DGU-20A5, a Shimadzu diode array detector

SPD-M20A, a Shimadzu column oven CTO-20, a Berthold radioflow detector LB513 with a

200 µL-liquid cell Z-200-6M, a LB5036 pump for supplementing liquid scintillation cocktail

at 3 mL/min (Berthold AG, Regensdorf, Switzerland), and a Shimadzu autosampler model

Accepted Manuscript

SIL-30AC. Radiochemical data acquisition was done using the RadioStar software (version

5.0.12.4, Berthold AG, Regensdorf, Switzerland). For the chromatographic analysis of [

C]-

selexipag and its metabolites, two different HPLC gradients were used. The first 67-min

gradient used 20, 35, 40, 60 75, 95, 95, 20 and 20% of mobile phase B at 0, 5, 15, 40, 47, 57,

58, 59 and 67 min run time, respectively. The second 90-min gradient used 20, 35, 40, 60, 62,

95, 95, 20 and 20% phase B at 0, 5, 15, 40, 80, 81, 82, 83 and 90 min run time, respectively.

The chromatographic separation was achieved using a Phenomenex Gemini C18 column

(5 µm, 250 x 4.6 mm ID, 110 Å) combined with a Phenomenex security guard cartridge

Gemini C18 (4 x 3 mm ID) at 45 °C with a flow rate of 1.0 mL/min. Mobile phases consisted

of 10 mM ammonium formate adjusted to pH 4.1 with formic acid (phase A) and acetonitrile

containing 0.1 % formic acid (phase B). Serial UV detection in a wavelength range of 190-

500 nm and

C -radiochemical detection were performed.

[

C]-ACT-333679 incubations with human liver microsomes, recombinant cytochrome

P450 and UGT enzymes

[

C]-ACT-333679 was incubated with pooled liver microsomes BD Biosciences (Tokyo,

Japan) containing 1 mg/mL proteins, in a 100 mM phosphate buffer (pH=7.4, 37 °C) at a

final concentration of 10 and 100 µM. The reaction was initiated by addition of the pre-

warmed NADPH-regenerating system and incubation was continued for 60 min at 37°C. In

the inhibition experiments with [

C]-ACT-333679 and furafylline (20 µM, CYP1A2), 8-

methoxypsoralen (5 µM, CYP2A6), sertraline (10 µM, CYP2B6), quercetin (5 µM,

CYP2C8), sulfaphenazole (3 µM, CYP2C9), N-benzylnirvanol (5 µM, CYP2C19), quinidine

(1 µM, CYP2D6), diethyldithiocarbamate (100 µM, CYP2E1) and ketoconazole (1 µM,

CYP3A4/5), each specific inhibitor methanol solution was placed into a polypropylene tube

and evaporated to dryness under a stream of nitrogen gas. Diethyldithiocarbamate aqueous

solution was used without dryness. The buffer and microsomes were added to the specific

Accepted Manuscript

inhibitor and stirred using a vortex mixer. The mixture was mixed with the test substance

solution and pre-incubated at 37°C for 5 min. NADPH regenerating system cofactor solution

was added to the mixture to start the reaction and incubated at 37°C for 1 h. As negative

controls, a NADPH regenerating system non-added group and specific inhibitor non-added

group were treated in the same way. Incubations of [

C]-ACT-333679 with recombinant

human P450-expressing microsomes CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1 and

CYP3A4 (50 pmol/ml) were performed at a single concentration of 10 µM and 100 µM in

100 mM phosphate buffer (pH 7.4, 37 °C) in duplicate. Similarly, [

C]-ACT-333679 was

incubated with recombinant UGT enzymes: UGT1A1, 1A3, 1A4, 1A6, 1A7, 1A8, 1A9,

1A10, 2B4, 2B7, 2B15 and UGT2B17 (1 mg proteins/mL) in 100 mM phosphate buffer (pH

6.0, 37 °C) in duplicate. The P450 reactions were initiated by adding the NADPH-

regenerating system in the pre-incubated mixture and incubation was continued for 1 h at

37°C. Control incubations in the absence of the NADPH-regenerating system, as well as

incubations with control microsomes without P450 enzymes at 1 mg/mL protein

concentration were performed in parallel under otherwise identical conditions. All reactions

were terminated by addition of two volumes equivalent of ice-cold methanol. The samples

were centrifuged at 22000 g for 5 min at 4°C and the procedure was repeated once with the

pellet. Supernatants were combined, evaporated to dryness and reconstituted in 350 μL of the

HPLC initial mobile phase. A 50 µL-aliquot of the supernatant was submitted to HPLC

analysis. The UGT reactions were initiated by adding a UDPGA cofactor solution (2 mM

UDPGA, 50 g /mL alamethicin, 5 mM D-saccharic acid 1,4-lactone monohydrate, and 10

mM MgCl

as final concentrations) in the pre-incubated mixture and incubation was

continued for 30 min at 37°C. The reaction was terminated by addition of two volumes

equivalent of ice-cold methanol containing hydrochloric acid. The samples were centrifuged

at 22000 g for 5 min at 4°C and the procedure was repeated once with the pellet.

Accepted Manuscript

Supernatants were combined, evaporated to dryness and reconstituted in 350 μL of the HPLC

initial mobile phase. A 50 µL-aliquot of the supernatant was submitted to HPLC analysis.

Control incubations in the absence of the UDPGA cofactor solution, as well as incubations

with control microsomes without UGT enzymes at the highest protein concentration were

performed in duplicate in parallel under otherwise identical conditions.

Analytical method for ACT-333679 incubations

Metabolic profiles were analysed using HPLC coupled with radioactivity detection. The

analytical system consisted of a delivery unit LC-10AD (Shimadzu, Kyoto, Japan), a

Shimadzu system controller SCL-10A, a Shimadzu UV-VIS detector SPD-10A, a Shimadzu

column oven CTO-10AS, a Shimadzu radioactivity detector (RAD) 525TR and a Shimadzu

autosampler model SIL-HTC. Radiochemical data acquisition was done using ChemStation

(Agilent) and the Flo-One software package version 3.65 built in the RAD. The

chromatographic separation of [

C]-ACT-333679 and its metabolites was achieved using a

Chemicals Evaluation and Research Institute L-column ODS (5 μm, 4.6 × 150 mm) at 40°C

with a flow rate of 1.0 mL/min. Mobile phases consisted of 0.1% formic acid in water (phase

A) 0.1% formic acid in methanol ( phase B). The applied gradient method started at 0 min

with 60% mobile phase B increased to 90% at 15 min and maintained until 25 min, then back

to 60% at 25.01 min and maintained until 30 min. UV detection at a wavelength of 254 nm

and radiochemical detection were performed.

Inhibition of cytochrome P450 enzymes

The potential of selexipag and its metabolite ACT-333679 to inhibit cytochrome P450

enzymes was studied in vitro using human liver microsomes (Xenotech LLC, Nosan

corporation, Japan) containing 0.2 mg/mL proteins in a 100 mM phosphate buffer (pH=7.4,

37 °C).The reaction was initiated by addition of the NADPH-regenerating system containing

Accepted Manuscript

the glucose-6-phosphate dehydrogenase and continued for 2-30 min at 37 °C. The following

P450 isoform-specific marker transformations and the substrate concentrations were used:

phenacetin-O-deethylation (CYP1A2, 6 µM), coumarin 7-hydroxylation (CYP2A6, 0.2 µM),

bupropion hydroxylation (CYP2B6, 20 µM), paclitaxel 6-hydroxylation (CYP2C8, 2 µM),

diclofenac 4'-hydroxylation (CYP2C9, 1 µM), (S)-mephenytoin 4'-hydroxylation (CYP2C19,

4 µM), bufarolol 1'-hydroxylation (CYP2D6, 2 µM), chlorzoxazone 6-hydroxylation

(CYP2E1, 8 µM), and midazolam 4'-hydroxylation (1 µM) and testosterone 6-hydroxylation

(10 µM) for CYP3A4. The biotransformation was assessed by product formation. Metabolite

formation was assessed by LC/MS-MS using qualified methods, calibration curves and

quality control samples. The following positive controls were used: -naphtoflavone

(CYP1A2, 0.2 µM), 8-methoxypsoralen, (CYP2A6, 0.2 µM), ticlopidine (CYP2B6, 0.2 µM),

quercetin (CYP2C8, 25 µM), sulfaphenazole (CYP2C9, 5 µM), tranylcypromine (CYP2C19,

25 µM), quinidine (CYP2D6, 1 µM), diethyldithiocarbamate (CYP2E1, 20 µM),

ketoconazole (CYP3A4, 0.5 µM). Selexipag and ACT-333679 were tested up to 50 µM in

duplicate. Samples were analysed using LC-MS/MS except for CYP2A6 (HPLC). The IC

values were calculated by Pharmacodynamic model 103 (Inhibitory effect E

max

model) with a

pharmacokinetic analysis software product, WinNonlin Version 5.2 (Pharsight Corporation).

The K

values were calculated by plotting (Dixon plot) the reciprocal metabolite formation

rates on the y-axis and test substance concentrations on the x-axis, followed by the analysis

with appropriate model formula using WinNonlin Ver. 5.2.1 (Pharsight). Time-dependent

inhibition was assessed on CYP2C8, CYP2C9, CYP2D6, and CYP3A4 using human liver

microsomes and comparing IC

values obtained with and without a pre-incubation period.

The so-called IC

shift is a marker of the extent of change in enzyme activity during the pre-

incubation period.

Accepted Manuscript

Induction of cytochrome P450 enzymes

Changes in CYP1A2, CYP2C9, CYP2B6, and CYP3A4 mRNA expression were assessed in

at least three different batches of freshly isolated or cryopreserved human hepatocytes,

obtained from Biopredic International (Rennes, France) and Bioreclamation IVT (Neuss,

Germany), respectively. Hepatocytes were incubated in triplicate wells with either DMSO

(control) or selexipag or ACT-333679 at concentrations ranging from 0.1 to 100 µM. The

positive controls were rifampicin for CYP3A4 and CYP2C9 (0.1-50 µM), phenobarbital for

CYP2B6 (100-3000 µM) and omeprazole for CYP1A2 (1-100 µM). Compound solutions were

prepared by diluting the respective DMSO stock solutions with cell culture medium by a factor of

1000, i.e. with 0.1% of DMSO in the incubation medium.. All incubations were done for 68 h

using cell culture media with 0.2% bovine serum albumin. The hepatocytes lysates were used

for total RNA extraction with a Qiagen Rneasy micro kit (Qiagen AG, Switzerland). DNase

(recombinant human desoxyribonuclease purchased from Qiagen) treatment was applied

during the RNA extraction process, to avoid contamination by genomic DNA.

Complementary DNA (cDNA) was synthesized using 50 ng of total RNA as template with

the Taqman reverse transcription reagents kit from Life Technologies Europe B.V. (Zug,

Switzerland) and MultiscribeTM reverse transcriptase at 1.25 IU/μL, random hexamers at 2.5

μM, and RNase inhibitor at 0.4 IU/μL. PCR was performed using the Taqman PCR

Mastermix from Life Technologies. CYP3A4, CYP2C9, CYP1A2, CYP2B6 and 18S

transcript levels were quantified with a StepOnePlus sequence detection system (Life

Technologies) according to the manufacturer’s protocol. Relative transcript levels in

compound-treated cells and DMSO-treated controls were determined using the comparative

Ct (cycle threshold) method (ΔΔCt method). The amount of target, normalized to an

endogenous control (18S) and relative to a calibrator (control sample in the experiment, i.e.

DMSO-treated hepatocytes) is given by 2-ΔΔCt. The fold-induction for each gene and

compound concentration was calculated using the relative transcript level ratio over the one

Accepted Manuscript

obtained for DMSO-treated cells. Statistical analysis and curve fitting for EC

determinations were performed using the GraphPad Prism software (version 5, GraphPad

Software Inc., San Diego, USA). For the construction of the concentration-response curves, a

three-parameter equation (sigmoid) was used Y = Bottom + (Top-Bottom)/(1 + 10((log EC50

- X), in which X equals the logarithm of concentration, and Y equals the fold-induction.

Inhibition of drug transporters

The effect of selexipag and ACT-333679 on the activity of the multidrug resistance protein

MDR1 (ABCB1, P-gp) was investigated using MDR1-overexpressing MDCK-II cells

(Netherland Cancer Institute, Amsterdam, Netherlands). Tritiated digoxin (1 M) and

rhodamine 123 (10 M) were used as substrates. Inhibition of P-gp by selexipag and ACT-

333679 were tested at concentrations up to 10 M. Experiments with digoxin were analysed

using a liquid scintillation analyser with luminescence correction and on-line quenching

correction by means of an internal standard (LSC) and those with rhodamine 123 by

spectrofluorimetry. Verapamil was used as a control inhibitor (100 µM). Apparent

permeability coefficients (P

app

) and net secretory flux were calculated as previously described

(Fenner et al., 2009). Inhibition of the efflux transporters BCRP, BSEP and MRP2 was

investigated in transporter-overexpressing Sf9 vesicles obtained from SOLVO Biotechnology

(Budaörs, Hungary). The tritiated model substrates methotrexate (100 M), taurocholic acid

(5 M) and estradiol-17-β-glucuronide (50 M) were used for BCRP, BSEP and MRP2

together for selexipag and ACT-333679 up to 1000 M. Control inhibitors were Ko143 (0.3-

2 µM), cyclosporine A (0.5-5 µM) and MK571 (1-100 µM), respectively. Samples were

analysed by LSC. Inhibition of the uptake transporters OATP1B1, OATP1B3 and OAT1 by

selexipag and ACT-333679 was investigated in Chinese hamster ovary (CHO) cells

overexpressing the respective transporters. The OATP1B1/1B3 overexpressing cells were

Accepted Manuscript

kindly provided by Professor B. Stieger (University of Zurich, Switzerland) and the OAT-1

cells were from SOLVO Biotechnology (Budaörs, Hungary). The tritiated model substrates

atorvastatin (0.5 M), taurocholic acid (5 M) and p-aminohippuric acid (1 M) were used

together with selexipag and ACT-333679 up to 100 µM and 1000 M. The positive controls

used were cyclosporin A (OATP1B1/1B3, 5 µM) and probenecid (OAT1, 5-100 µM).

Samples were analysed by LSC. Inhibition of the uptake transporters OAT3, OCT1 and

OCT2 and the efflux transporters MATE1 and MATE2K was investigated in transporter-

overexpressing human embryonic kidney (HEK293) cells, all from Absorption Systems LP

(Exton, USA). The respective model substrates and control inhibitors were: furosemide (5

M) and probenecid (100 µM); 1-methyl-4-phenylpyridinium iodide (MPP, 5 µM, for OCT1

and OCT2) and repaglinide (600 µM) and imipramine (300 µM) ; metformin (50 M) and

cimetidine (100 µM); 4-(4-(dimethylamino)-styryl)-N-methyl-pyridinium (ASP, 1 M).

Selexipag and ACT-333679 were tested at concentrations up to 100 M. Samples were

analysed using LC-MS/MS. For each transporter studied, the respective substrate incubation

period was selected so that transport was linear over time. All incubations were performed

either in duplicate or in triplicate in one or two independent experiments. Cellular and

vesicular uptake rates were normalized to total protein content and incubation. Cellular net

uptake rates were calculated as the difference in the uptake rate of the overexpressing cells

and corresponding wild-type cells. For the studies with BCRP, MRP2 and BSEP

overexpressing vesicles, transporter-mediated net uptake rates were calculated as the

difference of the values obtained in the absence and presence of adenosine triphosphate.

All data were evaluated by plotting the inhibitor concentration (logarithmic scale) against the

net uptake rate. IC50 values were determined from the plot by non-linear regression using he

GraphPad Prism software package (version 5.0, GraphPad Software Inc., La Jolla, CA) and

the following equation: Y = ((Top/(1+ (x/IC

)

) + Bottom, where y is the net uptake rate, x is

Accepted Manuscript

the inhibitor concentration (µM), s is the slope at the point of inversion, and Top and Bottom

the maximum and minimum values for the cellular uptake rate or % of control.

In vitro binding of selexipag and ACT-333679

The binding of selexipag and/or ACT-333679 was determined in human plasma, human liver

microsomes, and in the incubation media used in the in vitro perpetrator investigations on

cytochromes P450 and drug transporters. Human plasma containing EDTA and prepared at

Covance Laboratories Ltd (Harrogate, England) was fortified with [

C]-selexipag or [

C]-

ACT-333679 in a concentration range of 0.1-50 g/mL (0.2-100 nM or 0.24-120 nM,

respectively) . Equilibrium dialysis was performed using a Dianorm equilibrium dialyser

(Pierce and Warriner, Chester, UK) at 37°C against 10 mM phosphate-buffered saline (pH

7.4) for 4 h, the time at which equilibrium was reached. Following dialysis, aliquots were

taken from the donor and buffer compartment, and total radioactivity was quantified by LSC.

The binding in human liver microsomes and incubations media were determined by

equilibrium analysis using a Pierce Rapid Equilibrium Dialysis (RED) device (Thermo Fisher

Scientific, Lausanne, Switzerland) and non-labelled selexipag or ACT-333679. Incubations

were carried out on a shaker at 37°C under 5% CO

atmosphere for 4 h. The donor

compartment was either 1 mg/mL human liver microsomes, vesicles or cell culture medium.

Donor and receiver (containing phosphate buffer, pH 7.4) were analysed by LC-MS/MS and

the free fraction and recovery were calculated.

Estimation of [I, gut] for selexipag

[I, gut], i.e. calculated gut concentration, was estimated following the current European

medicines agency (EMA) and the food and drug administration (FDA) guidelines for drug-

drug interactions (EMA, 2012; FDA, 2012) for cytochrome P450 enzymes, and as described

Accepted Manuscript

previously for P-gp/MDR1 (Agarwal S, Arya V and Zhang L, 2012). The equation used was

[I, gut] = f

 k

 Dose/Q

, where f

is the fraction absorbed, k

the absorption rate constant,

and Q

the blood flow through the enterocytes. The absorption rate constant was assumed to

be 1 and for the blood flow through the enterocytes, a value of 300 mL/min was used. The

fraction absorbed was determined using Phoenix WinNonlin v.6 (Certara, Princeton, NJ,

USA) based on observed plasma concentration time profiles of selexipag after oral dosing

using 1-compartamental modelling (Kaufmann et al., 2017).

Results

The metabolism and excretion of selexipag has been investigated in human after oral dosing

C-labeled drug and sampling of urine and faeces over time intervals up to 168 h to

determine the total radioactivity eliminated and allow the analysis of metabolites (Kaufmann

et al., 2012). Total radioactivity was eliminated primarily in the faeces, accounting for 92.7%

of the administered dose with 87.6% of the administered dose recovered within 96 h. Renal

excretion of drug-related material accounted for 11.9% of the dose. Overall, total

radioactivity recovered in urine and faeces represented 104.6% of the administered dose.

To determine the nature of the metabolites eliminated via the faeces and due to the low

absolute radioactivity contained in individual samples, those extracts with the highest

radioactivity were selected for further analysis. Metabolic patterns were recorded using

HPLC coupled to radiodetection. A representative chromatogram of a faecal sample collected

during a 24-48 h interval from a healthy subject is depicted in Figure 2A. Besides

ACT-333679, the main product P10 was observed, followed by P18. Selexipag was absent

from all faecal samples analysed. The identity of metabolites for which synthetic standards

were available (P10 and P18) was confirmed by comparing their chromatographic retention

times and MS/MS spectra.. [

C]-ACT-333679 and ACT-333679 were first incubated with

recombinant human CYP2C8 and each putative radiolabeled and unlabeled metabolite P10

Accepted Manuscript

isolatedChromatographic retention times of these isolated metabolites were identical to the

one of metabolite P10 in human faecal samples. The key fragment ions generated were m/z

394, 378, 318, 276 and 249 and were in agreement with the fragmentation pathways of the

respective authentic standard (Figure 2B-C). P10 was identified as a product of

hydroxylation at the phenyl ring of ACT-333679. Similarly, isolated radiolabeled and

authentic standard P10 were further incubated with recombinant CYP2C8 and each putative

radiolabeled and unlabeled metabolite P18 isolated. The resulting product, which retention

time was identical to that of P18, was isolated by HPLC and unequivocally identified by

comparison of its HPLC retention time and mass fragmentation pattern with those of an

authentic standard (Figure 2, D-E). P18 was identified as a hydroxylation product of P10. In

conclusion, P10 was identified as a product of hydroxylation at the phenyl ring of ACT-

333679 and was further hydroxylated at the isopropyl group to form P18.

Due to the low total radioactivity content in the urine and plasma samples from the

radioactive study in man, no metabolite identification studies could be performed on them.

Therefore, time-pooled plasma samples from a separate clinical phase I study following

multiple doses of 1600 µg and 1800 µg b.i.d (Bruderer et al., 2014) were used to determine

the nature and the amount of circulating metabolites by LC-MS/MS analysis. For the

metabolite identification work, human plasma samples from the 1800 µg dose group were

initially pooled per time point. In addition to the already known metabolite ACT-333679,

metabolites P4, P10, P11, P12, P13, P14, P34 and P35 were identified. The proposed

structures of these metabolites were confirmed by comparison with chemical references

which displayed identical retention time and MS fragmentation pattern. The quantification of

these metabolites was carried out using LC-MS/MS and an additional set of human plasma

samples from the 1600 µg dose group pooled per time point at: pre-dose, 2 h, 4 h, 6 h and 12

h post-dose using known standards. The relative abundance of selexipag and its metabolites

Accepted Manuscript

in human plasma, as well as the determined and exact theoretical mass values for the

metabolites are summarized in Table 1. Based on the AUC

0-last

values, selexipag and

ACT-333679 were the most abundant entities representing 14% and 78% of drug-related

material, while none of the other metabolites (P4, P10-P14, P34 and P35) exceeded 3% of

drug-related material.

The identification of enzymes involved in the biotransformation of selexipag was performed

in vitro using various biological preparations of human origin, including plasma, hepatocytes,

liver microsomes and also recombinant phase I and phase II enzymes.

To determine the nature of the hydrolytic enzymes involved in the formation of ACT-333679

from selexipag, human liver microsomes were used in the presence and in the absence of

hydrolase inhibitors. BNPP and PMSF were used as carboxylesterase (CES) inhibitors,

EDTA as a paraoxonase inhibitor and eserine as an acetylcholine esterase inhibitor.

Hydrolysis of selexipag leading to the formation of ACT-333679 was absent from

incubations in human plasma. The effect of the various inhibitors on the formation rate of

ACT-333679 from selexipag in human liver microsomes is shown in Figure 3. Incubations in

the absence of the NADPH-regenerating system yielded significant amounts of ACT-333679,

i.e. with a formation rate between 13-19 pmol/min/mg microsomal protein. Hydrolysis, i.e.

ACT-333679 formation, was suppressed in the presence of the CES inhibitors BNPP and

PMSF, whereas the paraoxonase and acetylcholine esterase inhibitors EDTA and eserine had

little or no effect.

The human P450 enzymes catalyzing the metabolism of selexipag and ACT-333679 were

investigated as follows: two complementary approaches were pursued, i.e., incubation with

liver microsomes in the absence or presence of chemical inhibitors for the various P450

isoforms, and incubation with recombinant P450 enzymes. Parallel sets of experiments were

performed with and without the esterase inhibitor BNPP to suppress hydrolysis of selexipag

Accepted Manuscript

and its metabolites by microsomal enzymes. Incubations with human hepatocytes were also

performed. Metabolic profiles were recorded using HPLC coupled to radiodetection.

Representative radiochromatograms of incubations with human liver microsomes,

recombinant CYP2C8, CYP3A4, UGT1A3 or UGT2B7 are illustrated in Figures 4-6.

Metabolites generated after incubations in human liver microsomes or hepatocytes are

summarized in Table 2.

Formation of metabolites P8, P10, P14, P36, P39, P40 and ACT-333679 was observed after

the incubation of [

C]-selexipag with liver microsomes representing altogether about 30% of

the respective radiochromatograms. ACT-333679 was the most prominent metabolite. Only

ACT-333679 and P10 were affected by the presence of BNPP suggesting a role of

carboxylesterase in their formation. The CYP2C8 inhibitor montelukast reduced P8

formation and completely inhibited P39 formation. Montelukast reduced by half the

formation of P10. The CYP3A4 inhibitors azamulin and ketoconazole led to a reduction in

P40 formation. Metabolite P36 formation was completely inhibited in the presence of

azamulin whereas its formation was only reduced by ketoconazole. Metabolite P14 was not

detected in experiments with azamulin. Inhibition of P14 formation by ketoconazole resulted

only in a partial reduction. The formation of P14 was not detected in presence of the

CYP1A2 selective inhibitor furafylline. Incubations in the combined presence of

montelukast, ketoconazole and BNPP were also investigated in order to assess the overall

contribution of CYP2C8, CYP3A4 and carboxylesterase in selexipag metabolism. The

formation of metabolites P10, P14, P36, P39 and P40 was completely inhibited by

combination of the three inhibitors. P8 formation was reduced by 60%. When [

C]-selexipag

was incubated with recombinant P450 enzymes, turnover was mostly observed with CYP2C8

and CYP3A4. The formation of P8 and P39 was only seen with CYP2C8. Formation of P14,

P36, P40 and P41 was observed in the presence of CYP3A4. The formation of P41 was

Accepted Manuscript

decreased in the combined presence of CYP3A4 and CYP1A2. Conversely, P14 formation in

presence of CYP3A4 was increased in the presence of the CYP3A4/CYP1A2 combination.

When [

C]-ACT-333679 was incubated with human liver microsomes, P10 was by far the

main metabolite observed and its formation was reduced in the presence of the CYP2C8

inhibitor quercetin. The incubation of [

C]-ACT-333679 with recombinant CYP2C8 lead to

the abundant formation of P10, which was not observed with any other recombinant P450

tested. Also, using recombinant CYP3A4 and [

C]-ACT-333679, several products were

observed that were not structurally identified. Since P11, the acylglucuronide of ACT-

333679, was identified in small amounts in human plasma, [

C]-ACT-333679 was incubated

with a battery of recombinant UGT enzymes. The formation of P11 was mostly observed in

the presence of UGT1A3 followed by UGT2B7.

The binding of selexipag and ACT-333679 to human plasma proteins was investigated using

equilibrium dialysis. Mean plasma binding values of selexipag and ACT-333679 were 99.7 %

and 99.6 %, respectively.

The potential of selexipag and ACT-333679 to inhibit cytochrome P450 enzymes was studied

in vitro using human liver microsomes and P450 isoform-specific marker transformations.

Time-dependent inhibition was assessed for both compounds on CYP2C8, CYP2C9,

CYP2D6 and CYP3A4 using human liver microsomes. The results of the competitive

inhibition experiments and the derived free inhibition constant (K

) values are summarized in

Table 3. Selexipag only weakly affected most of the human P450 enzymes. IC

values were

around or above 15 M for CYP1A2, CYP2A6, CYP2B6, CYP2C19, CYP2D6, CYP2E1 and

CYP3A4. There was stronger inhibition of two members of the CYP2C family. K

values for

CYP2C8 and CYP2C9 were 0.6 M and 1.1 M. ACT-333679 was a less potent inhibitor.

With the exception of CYP2C8 and CYP2C9, IC

values for all P450 enzymes were above

15 M. K

values for CYP2C8 and CYP2C9 were 7.7 M and 10.5 M. Neither selexipag

Accepted Manuscript

nor ACT-333679 elicited a time-dependent inhibition of CYP2C8, CYP2C9, CYP2D6, or

CYP3A4.

The potential of selexipag and ACT-333679 to induce cytochrome P450 enzymes was studied

in vitro using human hepatocytes. Changes in CYP1A2, CYP2C9, CYP2B6, and CYP3A4

mRNA expression were assessed in different batches of primary human hepatocytes. The

results of these experiments and the derived free EC

values are summarized in Table 4.

Both compounds elicited concentration-dependent increases in CYP3A4, CYP2C9, and

CYP2B6 mRNA. Selexipag and ACT-333679 had no relevant effect on CYP1A2 mRNA.

The corresponding mean EC

values for selexipag on CYP3A4, CYP2C9 and CYP2B6

mRNA were 27 nM, 7 nM and 140 nM. ACT-333679 was generally less potent and the

values were 100 nM, 30 nM and 130 nM.

The potential of selexipag and ACT-333679 to inhibit drug transporters was assessed in cells

and membrane vesicles overexpressing the protein of interest. The results of these

experiments and the derived free IC

values are summarized in Table 5. Selexipag inhibited

the uptake transporters, OCT1 and OCT2, and the efflux transporters P-gp/MDR1, BSEP,

MATE1, MATE2K, and MRP2 with IC

values ranging from 8 µM to above 80 μM.

Stronger inhibition was observed on the uptake transporters OATP1B1, OATP1B3, OAT1,

and OAT3 with IC

values in the range of 1.0-1.7 µM. Selexipag showed a similar inhibition

of the efflux transporter BCRP with an IC

of 0.8 µM. ACT-333679 was generally less

potent. Except for OATP1B1, OATP1B3, BCRP, and OAT3, IC

values were higher than 10

µM. IC

values for OATP1B1, OATP1B3, BCRP, and OAT3 were in a range from 1.7-3.3

µM.

To assess the potential for selexipag to cause drug-drug interactions in the gastrointestinal

tract, [I,gut] was calculated. . [I, gut] after the maximal selexipag dose of 1600 µg was 81

ng/mL (Table 6).

Accepted Manuscript

Discussion

The metabolism pathways of selexipag, investigated in human after oral dosing and

completed by in vitro studies in various enzymatic systems, are depicted in Figure 7.

Selexipag undergoes several primary biotransformation reactions: hydrolysis of the

sulfonamide leads to the formation of the pharmacologically active ACT-333679. Aromatic

hydroxylation of the pyrazine ring produces P36, while hydroxylation at the phenyl ring gives

rise to P8. Hydroxylation at the isopropyl group produces P39. Dealkylation of the

aminopyrazine resulting in the loss of the isopropyl moiety yields P40. An unusual ring

contraction of the pyrazine to an imidazole ring produces P35. Finally, another dealkylation

of the aminopyrazine resulting in the loss of the butoxy-N-(methanesulfonyl)acetamide

moiety yields P41. Most primary metabolites undergo subsequent hydrolysis to their

corresponding acids P4, P10, P12, P13 and P34. Further dealkylation of P41 leads to the

formation of P14. ACT-333679 is conjugated with glucuronic acid to the acyl-glucuronide

P11, and hydroxylated at the phenyl ring to form P10. P10 is then further oxidized at the

isopropyl group to yield P18.

The main metabolic pathway of selexipag is the formation of the active metabolite ACT-

333679. Except for ACT-333679, none of the circulating metabolites in human plasma

exceed 3% of the total drug-related material. The inhibition of ACT-333679 formation in

human liver preparations by the CES inhibitors as well as the absence of hydrolysis of

selexipag in human plasma (Li et al., 2005) indicate a role for hepatic CES enzymes. The

human P450 enzymes involved in selexipag metabolism have been identified. In addition to

ACT-333679, five primary metabolites of selexipag were formed. Metabolites P8 and P39

were formed by CYP2C8 whereas metabolites P36, P40 and P41 were formed by CYP3A4.

Metabolite P41 was N-dealkylated to the secondary metabolite P14 by CYP1A2. The

experiments on P450 enzymes involved in the metabolism of ACT-333679 focused on P10

Accepted Manuscript

formation as the main metabolite in human faeces. CYP2C8 was the only P450 isoform

catalyzing the aromatic hydroxylation to P10. CYP2C8 together with CYP3A4 were also

involved in the formation of several minor ACT-333679 metabolites. There was evidence

that CES contributed to the formation of the secondary metabolite P10 from metabolite P8.

UGT1A3 and UGT2B7 catalyzed the glucuronidation of ACT-333679 to the acyl

glucuronide P11.

The human pharmacokinetics of selexipag after oral dosing have been described in several

studies (Bruderer et al., 2014; Baldoni et al., 2015; Hoch et al., 2015). Selexipag is rapidly

absorbed and is hydrolysed to its active metabolite ACT-333679. Maximum observed plasma

concentrations of selexipag and ACT-333679 are reached within 1–3 h and 3–4 h,

respectively. A study after intravenous dosing of 200 µg selexipag was performed to

determine its absolute bioavailability which was 49 % (Kaufmann et al., 2017). In healthy

subjects and in patients with PAH, plasma exposure to ACT-333679 at all doses and at

steady-state is approximately 3-to 4-fold higher than to the parent compound. In the study

after intravenous dosing of selexipag, the mean exposure ratio of ACT-333679 to selexipag

was approximately 1.3. These data indicate a pre-systemic formation of ACT-333679,

potentially by intestinal CES. Indeed, it is known that in man CES enzymes are expressed in

the liver and in the gut and exist as different isoenzymes. CES1 is mostly expressed in the

liver, while CES2 is present in the liver and the intestine (Yang et al., 2009). A pre-systemic

contribution of CES2 to the formation of ACT-333679 is therefore likely.

The potential of selexipag and ACT-333679 as perpetrators of metabolic- and/or transporter

driven drug interactions was studied in vitro. In order to connect the in vitro data to the

selexipag and ACT-333679 exposures observed in the clinic, two separate assessments were

done. The first one considered the potential perpetrator effects on P450 enzymes and

transporters in liver and kidney, the second one in the intestine. For the first assessment,

Accepted Manuscript

knowledge of total peak plasma concentrations of selexipag and ACT-333679 were used. At

the highest treatment dose of 1600 g, these were 19.8 ng/mL and 28.7 ng/mL, i.e., about

40 nM and 70 nM for selexipag and ACT-333679. The respective free concentrations

calculated after correction for plasma protein binding were 0.12 nM and 0.28 nM. The

strongest inhibitory effects for selexipag/ACT-333679 on P450 enzymes were observed on

CYP2C8 and CYP2C9 Table 3). Similarly, the strongest inductive effects for

selexipag/ACT-333679 were seen on CYP3A4 and CYP2C9 (Table 4). These in vitro

concentrations are several magnitudes higher than the free observed plasma concentrations

rendering such interactions unlikely to happen in the clinic. In support of this assessment, no

change in the exposure to S-warfarin (CYP2C9 substrate) or R-warfarin (CYP3A4 substrate)

was observed in a healthy subjects study where selexipag was dosed at 400 g b.i.d. after a

single dose of 20 mg warfarin (Bruderer et al., 2016). Finally, for the inhibition of uptake and

efflux transporters, the strongest inhibitory effects for selexipag and ACT-333679 were

observed on OATP1B1, OATP1B1B3, OAT1, OAT3 and BCRP and were between 0.8-1.7

µM and 1.7-20 µM (Table 5). For the same reasons as indicated above for P450 enzymes,

such interactions are unlikely to happen.

For the second assessment dealing with the potential inhibitory effect of selexipag on

intestinal P450 enzymes and transporters, the gut concentration was calculated according to

the current EMA and FDA guidelines (EMA, 2012; FDA, 2012), i.e. using calculated gut

concentrations [I,gut] and comparing them to the IC

for BCRP inhibition, or EC

for

CYP3A4 induction. The [I,gut] after the maximal selexipag dose of 1600 µg was calculated

to be 81 ng/mL (Table 6), which is much lower than the IC

for BCRP (400 ng/mL) and also

less than 0.1 times higher than the EC

for CYP3A4 induction (13 ng/mL), making it

unlikely that a drug interaction on the level of the gut via BCRP inhibition or CYP3A4

induction will occur. This assessment was followed-up and confirmed by a clinical study

Accepted Manuscript

investigating the effect of a 1600 µg oral dose of selexipag on the pharmacokinetics of the

sensitive CYP3A4 substrate midazolam that has a significant gut extraction (Juif PE et al.,

2017, submitted for publication). This study showed that exposure to midazolam and its

metabolite 1-hydroxymidazolam was not affected by co-administration of 1600 g selexipag.

In conclusion, the metabolism pathways of selexipag have been characterized in man and it

has been shown that selexipag has a low potential to inhibit or induce cytochrome P450 and

transporters.

Accepted Manuscript

References

Agarwal S, Arya V, Zhang L. (2013). Review of P-gp inhibition data in recently approved

new drug applications: utility of the proposed [I(1) ]/IC(50) and [I(2) ]/IC(50) criteria in the

P-gp decision tree. J Clin Pharmacol 53(2):228-33.

Baldoni D, Bruderer S, Muhsen N, et al. (2015). Bioequivalence of different dose-strength

tablets of selexipag, a selective prostacyclin receptor agonist, in a multiple-dose up-titration

study. Int J Clin Pharmacol Ther 53(9):788–798.

Bruderer S, Hurst N, Kaufmann P, et al. (2014). Multiple-dose up-titration study to evaluate

the safety, tolerability, pharmacokinetics, and pharmacodynamics of selexipag, an orally

available selective prostacyclin receptor agonist, in healthy subjects.

Pharmacology 94(3-4):148-56

Bruderer S, Okubo K, Mukai H, et al. (2016). Investigation of potential pharmacodynamic

and pharmacokinetic interactions between selexipag and warfarin in healthy male subjects.

Clin Ther 38(5):1228-1236.

Clapp LH, Gurung R. (2015). The mechanistic basis of prostacyclin and its stable analogues

in pulmonary arterial hypertension: Role of membrane versus nuclear receptors.

Prostaglandins Other Lipid Mediat 120:56–71.

Duggan ST, Keam SJ, Burness CB. Selexipag. (2017). A Review in Pulmonary Arterial

Hypertension. Am J Cardiovasc Drugs 17(1):73-80.

Accepted Manuscript

EMA. European medicines agency. Committee for Human Medicinal Products (CHMP)

(2012). Guideline on the Investigation of Drug Interactions, CPMP/EWP/560/Rev.1 Corr. 2.

Available at:

http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/07/WC5

00129606.pdf.

FDA. United States Department of Health and Human Services, Food and Drug

Administration, Center for Drug Evaluation and Research (CDER) (2012). Guidance for

Industry: Drug Interaction Studies - Study Design, Data Analysis, Implications for Dosing,

and Labeling Recommendations. Available at:

https://www.fda.gov/downloads/drugs/guidances/ucm292362.pdf.

Fenner KS, Troutman MD, Kempshall S, et al. (2009) Drug-drug interactions mediated

through P-glycoprotein: Clinical relevance and in vitro-in vivo correlation using digoxin as a

probe drug. Clin Pharmacol Ther 85:173-181.

Galiè N, Corris PA, Frost A, et al. (2013). Updated treatment algorithm of pulmonary arterial

hypertension. J Am Coll Cardiol 62 (25 Suppl):D60–72. 12

Galiè N, Humbert M, Vachiery JL, et al. (2015). ESC/ERS Guidelines for the diagnosis and

treatment of pulmonary hypertension: The Joint Task Force for the Diagnosis and Treatment

of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European

Respiratory Society (ERS): Endorsed by: Association for European Paediatric and Congenital

Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT).

(2016). Eur Heart J 37:67-119

Accepted Manuscript

Gatfield J, Menyhart K, Wanner D, et al. (2017). The selexipag active metabolite ACT-

333679 displays strong anti-contractile and anti-remodeling effects, but low b-arrestin

recruitment and desensitization potential. J Pharm Exp Therap (E-published, doi:

10.1124/jpet.116.239665).

Hardin, EA, Chin KM. (2016). Selexipag in the treatment of pulmonary arterial hypertension:

design, development, and therapy. Drug Des Develop Ther 10:3747-3754

Hoch M, Darpo B, Remenova T et al. (2015). A thorough QT study in the context of an

uptitration regimen with selexipag, a selective oral prostacyclin receptor agonist. Drug Des

Devel Ther 9:175–185

Humbert M, Ghofrani HA. (2016). The molecular targets of approved treatments for

pulmonary arterial hypertension. Thorax 71:73-83

Juif PE, Boehler M, Donazzolo Y, et al. (2017). A pharmacokinetic drug-drug interaction

study between selexipag and midazolam, a CYP3A4 substrate, in healthy male subjects. Eur J

Clin Pharmacol. doi: 10.1007/s00228-017-2282-7.

Kaufmann P, Hurst N, Astruc B, et al. (2017). Absolute oral bioavailability of selexipag, a

novel oral prostacyclin IP receptor agonist. Eur J Clin Pharmacol 73(2):151-156

Accepted Manuscript

Kaufmann P, Okubo K, Sidharta PN, et al. (2012). Investigation of the absorption,

metabolism, and excretion of 14C-selexipag following oral administration to healthy male

subjects. Clin Pharmacol Ther 91(suppl. 1):S85.

Li B, Sedlacek M, Manoharan I, et al. (2005). Butarylcholinesterase, paraoxonase and

albumin esterase, but not carboxylesterase, are present in human plasma. Biochem Pharmacol

70:1673-84.

Sardana M, Moll M, Farber HW. (2016). Pharmacokinetic drug evaluation of selexipag for

the treatment of pulmonary arterial hypertension. Expert Opin Drug Metab Toxicol

12(12):1513-1520.

Simonneau G, Gatzoulis MA, Adatia I, et al. Updated clinical classification of pulmonary

hypertension. (2013). J Am Coll Cardiol 62 (25 Suppl):D34–41

Taichman DB, Ornelas J, Chung L, et al. (2014). Pharmacologic therapy for pulmonary

arterial hypertension in adults: CHEST guideline and expert panel report. Chest 146(2):449–

475.

Tuder RM, Cool CD, Geraci MW, et al. (1999). Prostacyclin synthase expression is

decreased in lungs from patients with severe pulmonary hypertension. Am J Respir Crit Care

Med 159(6):1925–1932.

Accepted Manuscript

Yang D, Pearce RE, Wang X, et al. (2009). Human carboxylesterases HCE1 and HCE2:

ontogenic expression, inter-individual variability and differential hydrolysis of oseltamivir,

aspirin, deltamethrin and permethrin. Biochem Pharmacol 77:238-47.

Figure 1 Chemical structures of [

C]selexipag and [

C]ACT-333679. Asterisks denote

the position of the radiolabel, which is uniformly distributed throughout both phenyl

rings.

Accepted Manuscript

Figure 2 Identification of metabolites in faecal samples

Panel A: Radiochromatogram of a representative faecal sample after oral administration of [

C]selexipag (400

µg/45 µCi) to healthy volunteers. Retention times: P18 (7.5 min), P10 (13.8 min), ACT-333679 (19.4 min),

selexipag (21.9 min)

Accepted Manuscript

Panels B / C: 11 Product-ion mass spectrum of the authentic standard for P10 (B) and

metabolite P10 (C) at m/z 436, (M + H)

Accepted Manuscript

Panels D / E: 11 Product-ion mass spectrum of the authentic standard for P18 (D) and

metabolite P18 (E) at m/z 436, (M + H)

Accepted Manuscript

Figure 3 Effect of esterase inhibitors on the formation rate of ACT-333679 from

selexipag in human liver microsomes in the absence of NADPH

Accepted Manuscript

Figure 4 Metabolic profiles of [

C]selexipag and [

C]ACT-333679 in human liver

microsomes

Panel A: [

C]selexipag with human liver microsomes

Panel B: [

C]ACT-333679 with human liver microsomes

Accepted Manuscript

Figure 5 Metabolic profile of [

C]selexipag and/or [

C]ACT-333679 with recombinant

CYP2C8 and CYP3A4

Panel A: [

C]selexipag with recombinant CYP2C8

Panel B: [

C]ACT-333679 with CYP2C8

Accepted Manuscript

Panel C: [

C]selexipag with recombinant CYP3A4

Accepted Manuscript

Figure 6 Metabolic profile of [

C]ACT-333679 with recombinant UGT1A3 and

UGT2B7

Accepted Manuscript

Figure 7 Proposed metabolic pathways of selexipag in human

Accepted Manuscript

Table 1 Prevalence of selexipag and its metabolites in human

plasma

Compound

AUC

0-last

(pmol/mL*h)

Relative

AUC

(%)

Determined

mass

Exact theoretical

mass [M+H]

Change in elemental

composition

Selexipag

38.3

497.2221

497.2223

ACT-333679

213

420.2285

420.2287

- CH3NOS (77)

P11

2.9

0.6

596.2626

596.2608

- CH3NOS (77) +

C6H8O6 (glucuronide,

+ 176)

P10

2.6

0.6

436.2232

436.2236

- CH3NS (61)

P12

2.6

378.1817

378.1818

- C3H6 - CH3NS (119)

P14

2.2

0.5

248.1186

248.1188

- C3H6 - C7H13NO4S

P13

7.1

2.6

436.2234

436.2236

- CH3NS (61)

2.6

0.9

436.2235

436.2236

- CH3NS (61)

P34

0.3

0.1

408.2287

- C2H3NOS (89)

Pharmacokinetic parameters were estimated using Phoenix™ WinNonlin (version 6, Pharsight Corporation, Mountain View, USA). A

non-compartmental PK analysis was performed calculating AUC

0-last

. Results are expressed in pmol/mL*h in order to account for the

differences in molecular weight.

AUC

0-last

were calculated based on the concentrations in acidified plasma,

sum of selexipag and all

metabolites = 100%. P35 was present in trace amounts. Full scan LC-MS runs of the sample containing the metabolites and control

sample were processed using the background subtraction tool of Metworks 1.2

Accepted Manuscript

Table 2 Prevalence of selexipag and its main metabolites in vitro in human liver

preparations following incubations with [

C]selexipag or [

C]ACT-333679

[

C]Selexipag

Human hepatocytes

Human liver microsomes

selexipag

6.9

73 / 87

ACT-333679

14 / 3.5

2.0

6.9 / 5.7

P10

1.2 /-

P14

2.0

2.3 / 1.9

P36

0.9

P39

0.6

P40

1.6

[

C]ACT-333679

Human hepatocytes

Human liver microsomes

ACT-333679

P10

P12

n.a.

2.7

P13

5.3

2.6

P14

2.1

P18

2.4

Incubation time:

24 h and

1 h.

Incubation in presence of the carboxylesterase inhibitor BNPP. Numbers are expressed as percent

of total radioactivity in the respective chromatograms; -: absence of a metabolite in the incubation. n.a., data not available.

Accepted Manuscript

Table 3 K

values of selexipag and ACT-333679 on several human cytochrome P450 enzymes

[µM]

Selexipag

ACT-333679

total

free

total

free

CYP2C8

2.0

0.6

7.7

CYP2C9

3.5

1.1

10.5

= inhibition constant;

Free K

was obtained using the free fraction determined in human liver microsomes, i.e.

30% for selexipag and 70% for ACT-333679 (both, SD < 5%).

Table 4 Regulation of CYP3A4, CYP2C9, and CYP2B6 mRNA in human hepatocytes

, mean

Selexipag

ACT-333679

total [µM]

free [nM]

total [µM]

free [nM]

CYP3A4

3.0 ± 0.6

8.4 ± 0.6

100

CYP2C9

0.8 ± 0.2

2.6 ± 0.5

CYP2B6

9.7 ± 0.3

140

10 ± 1.4

130

= concentration that induces 50% of the maximal response;

free EC

calculated using the fractions

determined in the different media used for the induction experiments, i.e. 0.8-1 % (SD < 5%) for selexipag and 1.1-

1.3% (SD < 5%) for ACT-333679.

Mean EC

values result from experiments in three independent hepatocyte

batches;

Mean EC

values result from one experiment in one batch.

Table 5 Inhibition of human transporter proteins by selexipag and ACT-333679

, [µM]

selexipag

ACT-333679

total

free

total

free

OATP1B1

2.4 ± 0.2

1.7

3.5 ± 0.3

2.6

OATP1B3

1.7 ± 0.2

1.2

4.1 ± 0.8

3.1

BCRP

1.9 ± 0.3

0.8

5.6 ± 0.9

2.2

OAT1

1.4 ± 0.2

1.0

25 ± 0.4

OAT3

1.7 ± 0.3

1.2

2.1 ± 0.6

1.7

= concentration that induces 50% of the maximal response. Mean values are given ± standard deviation (SD).

For P-gp/MDR1, OCT1/2, MATE1, MATE2K, MRP2 and BSEP IC

>10 µM.

Free IC

were calculated using

the mean free fractions determined for selexipag and ACT-333679 in cellular assays, i.e. 70% and 80%,

respectively; free fractions were 40% for both compounds in vesicles. SD on all mean binding values are below 8

Accepted Manuscript

Table 6 Calculation of gut concentration and drug interaction potential of selexipag

BCRP

CYP3A4 induction

fraction absorbed

absorption rate constant, ka (/min)

0.015

Qent (mL/min)

300

I,gut = Fa  ka  dose/Qent (g/mL)

0.081

ka (geometric mean) was determined by 1-compartamental PK modeling of the observed clinical data after oral

selexipag dosing (Kaufmann et al., 2017);

using a dose of 1600 g and Qent of 300 mL/min

Accepted Manuscript